9 #ifndef __IPLOWRANKUPDATESYMMATRIX_HPP__
10 #define __IPLOWRANKUPDATESYMMATRIX_HPP__
20 class LowRankUpdateSymMatrixSpace;
113 const std::string& name,
115 const std::string& prefix)
const;
LowRankUpdateSymMatrix()
Default Constructor.
bool IsValid(const SmartPtr< U > &smart_ptr)
Number * x
Input: Starting point Output: Optimal solution.
LowRankUpdateSymMatrix * MakeNewLowRankUpdateSymMatrix() const
Method for creating a new matrix of this specific type.
SmartPtr< const VectorSpace > LowRankVectorSpace() const
void SetV(const MultiVectorMatrix &V)
Method for setting the positive low-rank update part.
SmartPtr< const VectorSpace > lowrank_vector_space_
Vector space for the space in which the low-rank approximation lives.
virtual bool HasValidNumbersImpl() const
Method for determining if all stored numbers are valid (i.e., no Inf or Nan).
SmartPtr< const MultiVectorMatrix > U_
Vector storing the negative low-rank update.
double Number
Type of all numbers.
void ObjectChanged()
Objects derived from TaggedObject MUST call this method every time their internal state changes to up...
virtual void ComputeRowAMaxImpl(Vector &rows_norms, bool init) const
Compute the max-norm of the rows in the matrix.
LowRankUpdateSymMatrixSpace(Index dim, SmartPtr< const Matrix > P_LowRank, SmartPtr< const VectorSpace > LowRankVectorSpace, bool reduced_diag)
Constructor, given the dimension of the matrix.
EJournalLevel
Print Level Enum.
~LowRankUpdateSymMatrix()
Destructor.
This is the base class for all derived symmetric matrix types.
Template class for Smart Pointers.
SmartPtr< const LowRankUpdateSymMatrixSpace > owner_space_
corresponding matrix space
virtual ~LowRankUpdateSymMatrixSpace()
Destructor.
SmartPtr< const Matrix > P_LowRank() const
Return the expansion matrix to lift the low-rank update to the higher-dimensional space...
virtual void MultVectorImpl(Number alpha, const Vector &x, Number beta, Vector &y) const
Matrix-vector multiply.
SmartPtr< const VectorSpace > LowRankVectorSpace() const
Return the vector space in with the low-rank update vectors live.
void operator=(const LowRankUpdateSymMatrix &)
Overloaded Equals Operator.
SmartPtr< const MultiVectorMatrix > GetU() const
Method for getting the negative low-rank update part.
virtual void PrintImpl(const Journalist &jnlst, EJournalLevel level, EJournalCategory category, const std::string &name, Index indent, const std::string &prefix) const
Print detailed information about the matrix.
Class for Matrices with few columns that consists of Vectors.
SmartPtr< const Matrix > P_LowRank() const
void SetDiag(const Vector &D)
Method for setting the diagonal elements (as a Vector).
int Index
Type of all indices of vectors, matrices etc.
This is the matrix space for LowRankUpdateSymMatrix.
bool reduced_diag_
Flag indicating whether the diagonal matrix is nonzero only in the space of V or in the full space...
void operator=(const LowRankUpdateSymMatrixSpace &)
Overloaded Equals Operator.
bool ReducedDiag() const
Flag indicating whether the diagonal term lives in the smaller space (from P_LowRank) or in the full ...
SymMatrixSpace base class, corresponding to the SymMatrix base class.
Class responsible for all message output.
void SetU(const MultiVectorMatrix &U)
Method for setting the negative low-rank update part.
virtual SymMatrix * MakeNewSymMatrix() const
Overloaded MakeNew method for the SymMatrixSpace base class.
SmartPtr< const Vector > GetDiag() const
Method for getting the diagonal elements.
SmartPtr< const Vector > D_
Vector storing the diagonal matrix D.
LowRankUpdateSymMatrixSpace()
Default Constructor.
SmartPtr< const MultiVectorMatrix > V_
Vector storing the positive low-rank update.
Class for symmetric matrices, represented as low-rank updates.
SmartPtr< const MultiVectorMatrix > GetV() const
Method for getting the positive low-rank update part.
EJournalCategory
Category Selection Enum.
SmartPtr< const Matrix > P_LowRank_
Expansion matrix to lift the low-rank approximation into a possibly higher-dimensional space...
virtual void ComputeColAMaxImpl(Vector &cols_norms, bool init) const
Since the matrix is symmetric, the row and column max norms are identical.